For me, it's a speed problem

Story: The popularity of Firefox is definitely waning!Total Replies: 28
Author Content
Ridcully

Jul 09, 2014
9:30 PM EDT
I use an old version of Chrome (ver 21) most of the time, and it's because later versions of Chrome refuse to install on openSUSE 11.4 (missing files or summat). Firefox (ver 24.6 and the latest available for openSUSE 11.4) is kept purely as a backup. So far, Chrome outperforms Firefox in speed of access and rendering. It's unfair to compare these old versions of course, but that's how I see it. Firefox is great, but it's a snail with a foot-clamp compared with Chrome - well, on my system anyway. Both browsers are equal in their ability to display data/movies/etc.

I'd love to see some later comparisons.
AwesomeTux

Jul 10, 2014
12:05 AM EDT
Ridcully, like the article says Firefox's "inability to really utilize multiple CPU cores is a huge performance bottleneck", that is likely Firefox's main performance problem. Consider this scenario, you have a quad-core system and you're using Chrome with 4 tabs open, each of those tabs is able to consume an entire CPU core, but in Firefox that's not the case because everything Firefox does is in a single process and obviously restricted to the computing power of one CPU core. The fact that Firefox without multiprocessing nearly performs as well as Chrome, I think it's safe to assume that when Mozilla is finished with their work on multiprocessing, Firefox will easily outperform Chrome.
Ridcully

Jul 10, 2014
1:43 AM EDT
Hopefully, you are correct. I have a very soft spot for FireFox. However I also note the final summation in the article and from which I assume you draw your enthusiastic conclusions:

Quoting:Firefox still has many problems, though. Its inability to really utilize multiple CPU cores is a huge performance bottleneck, especially for gaming. Once Google fixes the bugs in Chrome that cause it to choke-up and slow down, Firefox will once again be playing catch up. And that's the real cause of Firefox's decline. In the end, it's always playing catch up.


And I read that to imply that FireFox will remain slower than Chrome.....even after the multicore fix. I hope I am wrong.
AwesomeTux

Jul 10, 2014
2:04 AM EDT
No, I don't think that will be the case. You see, Chrome "choke[s]-up and slow[s] down" when playing Dead Trigger even with its multiprocessing, while Firefox performs great without multiprocessing. It would only make sense that Firefox would perform even better with multiprocessing, while Chrome is likely nearing it's peak. I see Firefox taking a lead in performance for a few months, and then Chrome taking the lead again soon after.

Standard procedure for Firefox, really... it's so sad :-[
Ridcully

Jul 10, 2014
4:39 AM EDT
"Watch this space" seems to be the only answer. Unfortunately, my feeling is that FireFox is now in the decline stage......and at least some of the decline seems to be self-inflicted. Even if it does vanish, FFox did remarkable work, and best of all, it allowed Windows users to see that there really IS a world outside IE.
jdixon

Jul 10, 2014
6:19 AM EDT
> "Watch this space" seems to be the only answer.

An organization that's busy purging itself of politically incorrect views doesn't tend to have much time for programming.

The loss of the Google money might be the best thing that could happen to them. It might force them to refocus on what's really important: The code they produce.
mbaehrlxer

Jul 10, 2014
11:19 AM EDT
i used chromium until i discovered that it triggered the freezes that i experienced in compiz, whereas firefox didn't do that. the main features i liked from chromium were process separation and private mode (private mode is really nice for development when you want to test a site from the perspective of a new visitor who has not been to that site before. much easier than having to clear cookies and all the time)

by the time i detected the freeze problem, firefox had caught up with private mode and i found other features lessening the need for process separation. the main reason for that was that i could easily kill resource-heavy websites without killing the whole browser. when low on resources i ended up with most tabs killed, loading only those in active use.

now with firefox i solved this with an extension that can unload tabs, manually, or automatically after a few hours of their last use. that works just as well for the resource needs. the only downside is that now i don't get to see which sites are the resource hogs.

there are many upsides to firefox though, for example it does not load all tabs at startup. and it has tab-groups which allow me to switch between tabsets, depending on what i want to work on.

i need to sit down and finish writing an article about this topic that i started some time ago. (it's mostly ready, just needs some polishing)

greetings, eMBee.
DrDubious

Jul 10, 2014
11:25 AM EDT
"An organization that's busy purging itself of politically incorrect views" [citation needed]

One unpopular (and obviously incapable, given his unwillingness to engage on the matter) CEO-candidate resigned that I know of. The last thing I remember hearing Eich "produce" was some potential deal with a proprietary-codec company for "remote-desktop"/videogame stuff - not exactly a wonderful thing for an organization supposed to be devoted to openly-participatory web. Before that, it was "blaming capitulation to proprietary codecs on Google not solving their problem for them" (see also: the FSF's dead "PlayOgg" campaign, which seemed to be abandoned when Google said they wouldn't be doing Ogg Theora on Youtube).

Besides this, I know of nobody else that has resigned (or been "purged"), nor of any ongoing campaigns to do so. Has there been some sort of pogrom at Mozilla that we haven't heard about?

Firefox's problem is Mozilla's corporate introversion. The bloat isn't in the browser, it's in the corporation. They talk amongst themselves inside and don't communicate *with* people outside anymore. (They "announce" occasionally and feebly, fluffily "market to" the outside, but it's all plainly corporate PR.) As a result they seem obsessed, like Eich, with "internet TV"/video and not much else, except perhaps marketing bare-minimal FirefoxOS to "developing" countries, on the assumption that if and when they solidify a hold on the market THERE before Google/Apple can compete at those price-levels, they'll *surely* have enough market-share to start paying attention to that "open web" thing again...

(There were also two board-members who resigned, but they seemed, from what little information got out, to be more bothered by Mozilla's corporate introversion too[wanting some fresh blood in the CEO chair to try to get Mozilla to return to its intended direction rather than an "insider" who'd just keep steering it off the metaphorical cliff], nothing to do with "purging itself of politically incorrect views").
jdixon

Jul 10, 2014
1:21 PM EDT
> One unpopular (and obviously incapable, given his unwillingness to engage on the matter) CEO-candidate resigned that I know of.

He wasn't a CEO-candidate, he was the CEO. And we'll never know whether he was capable or not.

> The last thing I remember hearing Eich "produce"...

Not a big fan of Javascript, I take it?

> The bloat isn't in the browser, it's in the corporation.

I agree, which is why I made the comment about the Google money.

> Besides this, I know of nobody else that has resigned (or been "purged"), nor of any ongoing campaigns to do so.

Would you expect to? Things like CEO's being forced out make the news. Normal level folks don't.
AwesomeTux

Jul 11, 2014
8:14 AM EDT
jdixon, Eich was definitely not a capable CEO. At a bare minimum, a CEO needs to be able to take pressure from the public, to take an organization in a decisive direction and stand strongly behind that decision. Eich wasn't even able to take public (and peer) pressure in regards to his own personal political decisions. Remember, Eich wasn't forced out, he willingly stepped down.

If he were a capable CEO he would have voiced his opinion and stance plainly as "I am religious/homophobic, and therefore I don't and will never support gay marriage, but I still am your CEO. Deal with it." Which would have caused Firefox developers and many websites to continue to boycott Mozilla. It's logical that that's not what he wanted to happen, so he stepped down. His first and last capable action as CEO.
jdixon

Jul 11, 2014
8:39 AM EDT
> Remember, Eich wasn't forced out, he willingly stepped down.

I disagree. I think that If he hadn't stepped down, he would have been fired by the board. But that's something that can't be proven either way.

I agree that he didn't handle the matter appropriately, but if he didn't have the board's backing, it would have made no difference if he had.

> His first and last capable action as CEO.

Since that was essentially his only action as a CEO, how can you argue that he wasn't capable? :)
AwesomeTux

Jul 11, 2014
10:38 AM EDT
jdixon wrote:Since that was essentially his only action as a CEO, how can you argue that he wasn't capable? :)


Good point :-]
gus3

Jul 12, 2014
11:57 AM EDT
I thought about throwing some fuel on the fire, but...

It's a shame how badly this thread got de-railed.
DrDubious

Jul 12, 2014
9:52 PM EDT
>> The last thing I remember hearing Eich "produce"...

>Not a big fan of Javascript, I take it?


That was far from the "last" thing he produced. How long ago did he do that? (I'm pretty sure he didn't do it at Mozilla, either). Point is, he didn't SEEM to be "producing" much at Mozilla (though I can't tell if this is because he wasn't, or because Mozilla doesn't communicate with "outsiders" anymore so we never found out).

>> Besides this, I know of nobody else that has resigned (or been "purged"), nor of any ongoing campaigns to do so.

>Would you expect to? Things like CEO's being forced out make the news. Normal level folks don't.


Again, you are implying that there is some ongoing campaign to "purge itself of politically incorrect views" (and previously, not only that but that it is taking up so much of their attention that they can't get anything else done). Do you have some inside knowledge of such a ridiculous thing going on, or are you just extrapolating based on the one CEO-for-a-day (or whatever it was) badly fumbling and resigning?

I maintain that (to the extent that it's possible to tell from outside Mozilla's concrete shell of corporate accretion) Eich was one of the overly-comfortable insider executives steering the foundation off into "just-another-company"-land, and was a terrible choice for CEO for that reason. It just bugs me that people keep skipping right over that practical matter to complain that the only reason he didn't stay CEO was "political correctness".

(Supposedly, his anti-gay political donations had actually come out [ ha! no pun intended...] previously, but it really wasn't such a big deal when he wasn't supposed to be the public face of the whole foundation. For me personally, the specific cause he was donating to wasn't the important part but rather the fact that he was so out of touch that he felt thousands of dollars [supposedly there was more than one $1000 donation, can't recall if they were to different organizations or what] was pocket change small enough to be thrown down a hole for mere political spite. That's not the kind of person that is likely to appreciate how much of a barrier to entry the "poll-taxes" of proprietary patents and copyrights are to normal people to use the "open web", which explains why he seemed so unconcerned about proprietary codecs and disinterested in promoting legally-free media ["Google won't do it for me, I give up"]).

Yeah, I know, I get a little worked up over this one - Mozilla is the last major browser project that still seriously tries to care about that participatory-web stuff, and they have a serious cultural problem undermining them in that department...and it's not "political correctness". If they can't figure out what it really is and do something about it, they're going to wither and die, and the "open web" will merely be "whatever Google feels like bothering with in Chromium and doesn't feel like making a proprietary component of Chrome instead", aside from comparatively tiny projects like konqueror/rekonq. I hope most of us would agree that this would be a bad outcome.

Meanwhile, gus3, it's not that much of a derailment - we're still discussing causes of Firefox's apparent decline of popularity, but the "it's because everyone was a meanie to Eich!" red herring thing DOES seem to dominate these discussions when they come up anywhere. :-)
jdixon

Jul 12, 2014
9:58 PM EDT
> Do you have some inside knowledge of such a ridiculous thing going on, or are you just extrapolating

Just extrapolating based on past experience. YMMV.
NoDough

Jul 13, 2014
5:10 PM EDT
Quoting:...based on the one CEO-for-a-day (or whatever it was) badly fumbling and resigning?
Sorry, what fumble did he make as CEO?

Quoting:I maintain that (to the extent that it's possible to tell from outside Mozilla's concrete shell of corporate accretion) Eich was one of the overly-comfortable insider executives steering the foundation off into "just-another-company"-land, and was a terrible choice for CEO for that reason.
And the evidence for this is...???

Quoting:Mozilla is the last major browser project that still seriously tries to care about that participatory-web stuff, and they have a serious cultural problem undermining them in that department...and it's not "political correctness". If they can't figure out what it really is and do something about it, they're going to wither and die, and the "open web" will merely be "whatever Google feels like bothering with in Chromium and doesn't feel like making a proprietary component of Chrome instead", aside from comparatively tiny projects like konqueror/rekonq. I hope most of us would agree that this would be a bad outcome.
Fair enough. Being passionate about the future of FF is a good thing. Conflating that passion with the widely publicized reason for Eich's engineered departure is a bridge too far.
750

Jul 14, 2014
12:29 PM EDT
I wonder if Mozilla's problem is that of any geek org. They are caught between wanting to play with the latest and greatest (codecs etc) and staying true to principles.

Just observe the kind of hoopla RMS goes through to stay true to his principles, and for that he is branded a kook and a luddite by the "geekerati".

the problem for the web right now is that it is trading one monoculture (IE) for another (Webkit). And Google have taken the reins on the new monoculture, and are rampaging ahead of any standardization processes that are attempted.

For instance, you can upload or download whole folders in Chrome (and perhaps most Webkit based browsers). But this is not a established standard, and so Firefox can't.

But most third party developers don't care. And this function has been embraced by app devs on Android that want to offer a "cross-platform" way to handle file transfers (start the app, enter the url, drag and drop the folders).

Mozilla used to have a compatriot in the standards push, Opera. But now Opera has pretty much folded and adopted Webkit/Blink.
gus3

Jul 14, 2014
12:41 PM EDT
I for one do not brand RMS a "kook," and I defend him against those who do (as if he needs my help). Yes, I disagree with is world-view, but when it comes to IT, he is where the rest of the world should be. To those who say he's preaching the impossible, I point out that he proves it is completely possible.
Bob_Robertson

Jul 14, 2014
3:31 PM EDT
> To those who say he's preaching the impossible, I point out that he proves it is completely possible

The world owes a great debt of gratitude to RMS's principles and "cussed stubbornness".
krisum

Jul 18, 2014
5:08 AM EDT
@AwesomeTux

> Consider this scenario, you have a quad-core system and you're using Chrome with 4 tabs open, each of those tabs is able to consume an entire CPU core, but in Firefox that's not the case because everything Firefox does is in a single process and obviously restricted to the computing power of one CPU core.

This is incorrect. Firefox and nearly all other browsers use multiple threads which will in turn use all cores where required. Single or multiple process doesn't matter. In fact single process is better to conserve resources and generally one finds firefox in newer versions uses less resources than chrome. The only reason for going to multiple process is to avoid taking down or affecting the entire process due to a misbehaving plugin/addon/page. Other smaller advantage is that closing a page will clear all process related artifacts for multi-process but with threads some may remain if coding is not done carefully enough (memory leaks etc).
krisum

Jul 18, 2014
5:12 AM EDT
> i used chromium until i discovered that it triggered the freezes that i experienced in compiz

I experienced not-so-infrequent crashes of chrome/chromium pages regularly -- only silver-lining being that they did not crash the browser due to multi-process design. That and somehow I cannot stand the slow scrolling and UI of chrome/chromium for too long. FF has always been much smoother for me (followed closely by Opera).
AwesomeTux

Jul 18, 2014
6:45 AM EDT
krisum wrote:This is incorrect. Firefox and nearly all other browsers use multiple threads which will in turn use all cores where required. Single or multiple process doesn't matter. In fact single process is better to conserve resources and generally one finds firefox in newer versions uses less resources than chrome. The only reason for going to multiple process is to avoid taking down or affecting the entire process due to a misbehaving plugin/addon/page. Other smaller advantage is that closing a page will clear all process related artifacts for multi-process but with threads some may remain if coding is not done carefully enough (memory leaks etc).


This is all mostly incorrect. Firstly, only plugins in Firefox may be able to take advantage of multiple CPU cores. I see Flash do it quite often, as well as Java. Everything else in Firefox, from the rendering engine to the JavaScript engine are limited to the power of a single CPU core.

To showcase it, here is a page that will cause 100% CPU usage:

Quoting:

<!DOCTYPE html>
<html>
	<head>
		<meta charset="utf-8">
		<title>JavaScript CPU Eater</title>

<script type="text/javascript"> while (loop = true) { loop = true; } </script> </head> </html>


Copy and paste that in a new HTML file and open it in Chrome. Then check your system monitor / top. You will see one CPU core at 100%. Chrome will still be responsive, allowing you to open more tabs and load webpages and everything. If you open the same file again in a new tab, you will see two CPU cores at 100% and Chrome will remain responsive (assuming you have more than two cores).

Now open the same file in Firefox. Same thing, you will see one CPU core at 100%. However, Firefox will not respond to any clicks or keyboard input, nothing. Firefox will eventually have nothing else to do but ask you to kill the script.

If you load up that Dead Trigger game in Firefox and check your system monitor / top, you will notice it too does not cause more than one CPU core to peak.
Steven_Rosenber

Jul 18, 2014
3:00 PM EDT
It's vitally important to have a state-of-the-art Web browser not controlled by Google or Microsoft.
Bob_Robertson

Jul 18, 2014
3:57 PM EDT
Seconded, Steven.
jdixon

Jul 18, 2014
4:47 PM EDT
> It's vitally important to have a state-of-the-art Web browser not controlled by Google or Microsoft.

As long as the code is available, someone else can pick up the pieces after Mozilla gets done imploding. That's the advantage of FOSS.
krisum

Jul 20, 2014
12:50 PM EDT
> If you load up that Dead Trigger game in Firefox and check your system monitor / top, you will notice it too does not cause more than one CPU core to peak.

You have completely misinterpreted all that I wrote. For a *single* page both Firefox and Chrome will use single processor only. Multiple cores/processors will be made use by both Firefox and Chrome across pages.

The behaviour you describe simply says that Firefox is using the main thread for doing the processing of one of the pages. That design flaw can be easily fixed by spawning a separate thread for every page -- I guess it reuses the main thread for processing one of the pages and new threads as required for others.

This behaviour will not change by just going to multi-processor model. To change that behaviour the code needs to be rewritten to spawn a new process (or thread in current implementation) for even a single page leaving the main thread free to respond to other events.
AwesomeTux

Jul 21, 2014
1:37 AM EDT
krisum wrote:Multiple cores/processors will be made use by both Firefox and Chrome across pages.


This is wrong. Again I'll provide a code example...

Quoting:

<!DOCTYPE html>

<html> <head> <meta charset="utf-8"> <title>JavaScript CPU Eater</title>

<script type="text/javascript"> var lines = 1, char = 1;

window.onload = function() { document.body.innerHTML = '<b>' + lines + ':</b> '; lines++;

setInterval(function() { if (lines > 10) { lines = 1, char = 1; document.body.innerHTML = '<b>' + lines + ':</b> '; } else { if (char > 20) { document.body.innerHTML += '<br><b>' + lines + ':</b> '; char = 1; lines++; } else { document.body.innerHTML += '0'; char++; } } }, 0); }; </script> </head>

<body></body> </html>


Copy and paste that in a new HTML file and open it in Firefox, it should use a good bit of CPU. Now open as many new windows in Firefox as you like and have them all open the same file. You will see the CPU usage for Firefox go up but only ever take 100% of a single core. What I've noticed is that Firefox will switch which core it uses multiple times, but I think that is mainly handled by the OS.

Likewise, open the same file in Chrome, and again open many new windows at the same file. You will see Chrome not only use multiple core usage up to 100% each if enough windows are open, but you will also see Chrome become slower and slower, confirming my early remark on Firefox's single core design out performing Chrome despite Chrome's use of multiple cores.
krisum

Jul 21, 2014
4:39 AM EDT
@AwesomeTux

So your are correct that rendering uses a single thread. However, other tasks like video/audio encoding/decoding etc do use separate threads. This page has some details: https://blog.mozilla.org/products/2011/07/15/goals-for-multi...
AwesomeTux

Jul 21, 2014
5:23 AM EDT
@krisum

Interesting.

I can confirm that video/audio decoding does appear to utilize multiple CPU cores somewhat, however, it seems that the rendering of the video is still limited to the main process/thread. I say this because I watched these videos to test it out: https://archive.org/details/CloudsTimelapse6860s4kRes

And there is definitely usage of multiple cores going on, but only one core reaches above 20-30% -- as in one core goes to 100% the rest stay around 20-30%. And of course the videos are slow and choppy, whereas in Chrome they're rather smooth and each core goes near-equally above 50%.

Posting in this forum is limited to members of the group: [ForumMods, SITEADMINS, MEMBERS.]

Becoming a member of LXer is easy and free. Join Us!